Zepto

Zepto interview process (2026)

Blog / Zepto interview process (2026)

Zepto interview accounts vary by role, seniority and date. This guide connects eleven TechPrep exercises to the reports behind them and explains where the practice version differs from the reported task. It is a preparation guide, not a guaranteed interview schedule.Most named questions here come from 2024 candidate reports. A 2025 backend account describes DSA, low-level design and a hiring-manager discussion. A report titled October 2026 describes tree redistribution and AI-assisted debugging. Historical reports remain useful for practice, but confirm your own assessment format with the recruiter.

Coding and problem solving: The 2024 Bangalore account reports two coding tasks through BarRaiser. The 2025 backend account describes a one-hour DSA round with solution and complexity discussion. These are individual experiences, not fixed company-wide timings.
Low-level design and data modeling: Reported assignments include chess with class, API and database design and train booking with reusable seat segments. Clarify whether your round expects executable code, a design discussion or both.
Projects, fundamentals and role-specific exercises: The 2024 ecommerce account includes database and scalability follow-ups. Other reports describe AI-assisted debugging or a vanilla-JavaScript Todo List for a frontend role. Prepare for the role you applied for rather than assuming every candidate receives all of these rounds.

Use the nine DSA exercises to practise explaining algorithms and the two LLD exercises to practise state and invariants. The source and scope notes below identify historical reports and deliberate practice adaptations.

Algorithms and data structures: Nine exercises covering dynamic programming, binary search, stacks, hashing, greedy grouping, sorting and trees.
Low-level design: Chess rules and train-seat inventory, with clear boundaries between executable practice and broader design discussion.
Testing, databases and projects: Prepare to reason about failures, data models and implementation choices.
Preparation plan: Schedule the collection and general behavioral practice around your interview date.

1. Algorithms and data structures

The September 2024 Bangalore report names Minimum Cost for Tickets and Daily Temperatures. Practise choosing a dynamic-programming state for pass coverage and explaining why a monotonic stack finds the next warmer day.Another 2024 account reports the task behind Find First and Last Position of Element in Sorted Array. Test duplicate values, absent targets and boundary positions.The Gurgaon SDE-2 account explicitly links Count Good Meals and Largest Number. The second was discussed rather than coded. For pair counting, distinguish repeated values from distinct indices. For concatenation ordering, keep the result as a string and test shared prefixes and all-zero input.An August 2024 round-one account names Destination City and Rabbits in Forest. Explain the terminal-city invariant and why a partially observed color group still contributes its full required size.Distribute Coins in Binary Tree uses the same algorithm as the rider-redistribution story in the October 2026 report. A subtree can pass a surplus or deficit to its parent. Count transfers across each edge and check that the total number of coins equals the number of nodes.The historical 2022 on-campus report describes minimizing the largest adjacent-value difference along a graph path. Minimize Path Value is an equivalent algorithmic practice variant: it accepts weighted edges directly. Convert the reported node values into edge weights using abs(A[u] - A[v]). Do not replace the minimax objective with an additive shortest-path sum.

2. Low-level design

Multiplayer Chess Game provides executable practice for piece rules, turns, board state and move history. The reported chess assignment also asked for API and database design. Our exercise does not grade those additional parts. After coding, sketch how you would persist players and moves and expose game actions through an API.The train-booking account includes searching trains, checking seats, booking and cancellation, with the same seat reusable on non-overlapping journey segments. Train Seat Booking System makes that inventory model executable: availability is scoped by train and date, and bookings occupy the half-open interval between stop indices. Adjacent journeys may reuse a seat.The practice class runs sequentially. Its method signatures, validation results and booking-ID lifetime are TechPrep conventions. The original discussion also included fair concurrent booking and database modeling, which this grader does not assess. As a design follow-up, explain how transactions and conflict handling would protect the same inventory invariant.Other accounts discuss ecommerce coupons and expense-sharing CRUD. Those assignments are not interchangeable with a simple coupon-redemption counter or an equal-split balance tracker, so they are not presented as matching exercises in this collection.

3. Testing, databases and projects

The AI-debugging report describes reading an order-aggregation implementation, proposing tests in plain English, then using an AI tool to implement tests and investigate failures. A separate SDE-1 account also describes recording prompts and testing concurrent behavior. These are reported workflows, not a company-wide rule or an additional auto-graded exercise here.For general preparation, test empty batches, uneven partitions and more workers than items when those inputs are valid. Explain what each test proves before asking an assistant to write it. Treat concurrency claims as properties to demonstrate, not assumptions based on a lock appearing in the code.The ecommerce account discusses indexes, database choices and scalability. Choose one project and explain its data model, an important query and a failure you investigated. Distinguish transaction guarantees from application-level invariants.The frontend preview confirms a Todo List built without a framework. Only the public preview was available, so this guide does not invent additional frontend requirements or treat a React task as the same assignment.

4. Preparation plan

Start with an untimed pass through the nine DSA exercises. For each, write down the invariant, explain time and space costs, and test one boundary that could break an otherwise plausible solution. Revisit weak patterns before adding time pressure.Then work through the two LLD exercises. Trace state changes by hand and test failed operations as well as successful ones. For train booking, cover overlapping segments, adjacent segments, separate dates, cancellation and ownership checks.The roadmap also includes TechPrep’s general behavioral practice. Prepare examples of a technical trade-off, a production problem and a disagreement you helped resolve. These are preparation prompts, not a claimed list of Zepto behavioral questions. Ask the recruiter about the language, tools and expected depth for your specific role.

Conclusion

Use the Zepto roadmap to schedule the eleven exercises and general behavioral practice. Keep the distinction between reported questions, historical accounts and executable practice adaptations when deciding what to study next.